Portal   Forum   Members   Market   Gallery   Events

mac vs pc discussion

Discussion in 'Chit Chat' started by Zephor, Oct 26, 2001.

Thread Status:
Not open for further replies.
  1. mud

    mud Peasant

    Joined:
    Jun 14, 2000
    Messages:
    1,675
    Likes Received:
    0
    Market Rating:
    0
    The reason G4's keep up in the real world in pro stuff like photoshop is becasue the software that i made for those programs are ment ot run on a G4. That is how thye keep up.
     
  2. Zephor

    Zephor Peasant

    Joined:
    Mar 6, 2001
    Messages:
    1,263
    Likes Received:
    0
    Market Rating:
    0
    wasn't grizzly the one that said this post was useless
     
  3. Grizzly

    Grizzly Peasant

    Joined:
    Mar 3, 2001
    Messages:
    558
    Likes Received:
    0
    Market Rating:
    0
    I said it would drag on forever, not useless (well, maybe to Mac users).

    But since Jakeman is condoning this, I want to join the mudslinging.
     
  4. SpocKirk

    SpocKirk Forum Moderator and Donator

    Joined:
    Feb 28, 2000
    Messages:
    4,813
    Likes Received:
    25
    Market Rating:
    0
    what's to stop someone from just doing a timed countdown in assembly or machine language? That would be a highly accurate measure, since nothing is in the way, just using the CPU. Each would be as optimized as possible for each processor.
     
  5. Jakeman

    Jakeman MSC Founder and Donator

    Joined:
    Feb 26, 2000
    Messages:
    25,761
    Likes Received:
    27
    Market Rating:
    16
    i don't know enough about computers to make any supported arguments... and since i'm smarter than all of you... :p
     
  6. Haite

    Haite Forum Moderator

    Joined:
    Jan 29, 2001
    Messages:
    9,325
    Likes Received:
    34
    Market Rating:
    0
    I agree with Jakeman. I see specifically Grizzly discussing how it's not hard to keep the Athlon fead data because it operates on DDR. It may be 15% better then SDR, but how much better is 15% better (in numeric terms)? I mean, just because it's 15% faster doesn't mean that's sufficient.

    --Haite
     
  7. Grizzly

    Grizzly Peasant

    Joined:
    Mar 3, 2001
    Messages:
    558
    Likes Received:
    0
    Market Rating:
    0
    Read in context. I implore you, PLEASE!

    Let me make it as simple as possible

    1) Keeksy said that Athlon suffers from memory bandwidth

    2) I respond

    3) DDR = Double Data RAM

    4) SDRAM != Transfer and Receive at same time

    5) Therefore, DDR is faster than SDRAM, theoretically 2X

    6) Benchmarks cited: 15% improvement

    7) Why 15%?

    8) Because Athlon cannot use the extra bandwidth

    9) Why not?

    10) Because Athlon only designed to use SDRAM memory bandwidth

    11) Why, you ask?

    12) DDR can provide much more bandwidth

    13) Examples? GeForce256 SDR vs. GeForce256 DDR

    Now that I've gone through this very slowly, we can conclude that

    A) The Athlon cannot benefit much more from DDR memory

    B) Therefore, the Athlon does not suffer from memory bandwidth limitation

    C) Since the raw data calculations are faster than the G4 and ithe Athlon doesn't suffer from memory bandwidth bottlenecks, real world data would be faster

    Understand?

    www.anandtech.com does a very indepth job of explaining this, but I was afraid I would confuse you with their article :p
     
  8. SpocKirk

    SpocKirk Forum Moderator and Donator

    Joined:
    Feb 28, 2000
    Messages:
    4,813
    Likes Received:
    25
    Market Rating:
    0
    If you go through something slowly and use all uppercase (shouting), even people who don't understand english will understand what you are saying.
     
  9. Haite

    Haite Forum Moderator

    Joined:
    Jan 29, 2001
    Messages:
    9,325
    Likes Received:
    34
    Market Rating:
    0
    :rolleyes: I'll just take your word for it. Each processor has its place, and we're all Mac nazis here :D.

    --Haite
     
  10. Jakeman

    Jakeman MSC Founder and Donator

    Joined:
    Feb 26, 2000
    Messages:
    25,761
    Likes Received:
    27
    Market Rating:
    16
    wow, did you make that up by yourself? :rolleyes:

    sure i can make a statement like RISC is better than CISC, but i have nothing to back that statement except for heresay (sp?) from other people. there are an infinite number of factors that contribute to everything about computer performance. so shaddup! :)
     
  11. Wulf

    Wulf MSC Knight and Donator

    Joined:
    Jul 19, 2000
    Messages:
    4,856
    Likes Received:
    10
    Market Rating:
    0
    YEah.... welll my butt can proccess floating point calculations faster than both of 'em :)
     
  12. Elbereth

    Elbereth MSC Commander and Donator

    Joined:
    Jun 17, 2000
    Messages:
    3,356
    Likes Received:
    14
    Market Rating:
    0
    Time for a new title - Wulf "Floating point Butt"

    :D
     
  13. Zephor

    Zephor Peasant

    Joined:
    Mar 6, 2001
    Messages:
    1,263
    Likes Received:
    0
    Market Rating:
    0
    these threads always go off really well until the 3rd page
     
  14. Grizzly

    Grizzly Peasant

    Joined:
    Mar 3, 2001
    Messages:
    558
    Likes Received:
    0
    Market Rating:
    0
    Nope, there was an article on Anandtech on why the P4 would perform so well with Rambus and why the Athlon didn't perform so well with DDR.

    Found the article: http://www.anandtech.com/chipsets/showdoc.html?i=1518&p=7

    RDRAM doubles the memory bandwidth of SDRAM. That's because the P4 is bottlenecked by SDRAM and RDRAM provides the full bandwidth a P4 needs.

    And the other one: http://www.anandtech.com/chipsets/showdoc.html?i=1386&p=3

     
    Last edited: Oct 28, 2001
  15. Keeksy

    Keeksy Peasant

    Joined:
    Jun 15, 2000
    Messages:
    1,876
    Likes Received:
    0
    Market Rating:
    0
    Are you telling me the Athlon isn't bandwidth starved? Try increasing the memory and fronside bus frequency to 166MHz (remeber to change the CPU multipler, so your chip stays the same). If your Athlon was satisfied before, then you won't see any improvement. That's not what will happen however. Such an increase in bandwidth will give you an extra 10% or more. Continuing to increase the memory and fronside bus clocks will result in simular gains. Why does this happen? Because memory is a bottleneck. As that bottleneck becomes smaller, the Athlon gains speed. I know the nForce chipset offers 4.2GB of memory bandwidth, yet the 2.1GB KT266A is faster. I'm still right when you realize that that 4.2GB of data has to travel over a 2.1GB fronside bus. Now there's a botteneck. I guess nVidia only gave the nForce that much bandwidth so there intigrated Geforce2 MX would be happy. :p Frontside bus and memory should be equal. It makes things simplier.
     
    Last edited: Oct 28, 2001
  16. Grizzly

    Grizzly Peasant

    Joined:
    Mar 3, 2001
    Messages:
    558
    Likes Received:
    0
    Market Rating:
    0
    10% is nothing.

    http://www.anandtech.com/video/showdoc.html?i=1107&p=6

    At 1024x768, 32bit color, DDR shows a 80% increase improvement over SDR.

    That shows the true potential of DDR. The Athlon improves nowhere near 80%. As I say once again, the Athlon does not have memory bottlenecks.
     
  17. Grizzly

    Grizzly Peasant

    Joined:
    Mar 3, 2001
    Messages:
    558
    Likes Received:
    0
    Market Rating:
    0
    That makes absolutely no sense.

    I think you read the info wrong:

    from http://www.anandtech.com/chipsets/showdoc.html?i=1535&p=2
    Let me translate:

    1) The Athlon cannot use more than one DDR266 channel

    2) Since it cannot use more, more bandwidth does not help

    3) Since more bandwidth does not help, the Athlon does not suffer form bandwidth limitation
     
  18. Keeksy

    Keeksy Peasant

    Joined:
    Jun 15, 2000
    Messages:
    1,876
    Likes Received:
    0
    Market Rating:
    0
    Yeah that's only if you use a 266MHz frontside bus and increase memory bandwidth. Like I said before, you have to increase both, and what I said makes prefect sence. :p That artical probably doesn't consider this.
     
  19. Jakeman

    Jakeman MSC Founder and Donator

    Joined:
    Feb 26, 2000
    Messages:
    25,761
    Likes Received:
    27
    Market Rating:
    16
    i'm gonna make a hardware review section on my site and put stupid stuff supporting whatever claims i choose to make. that way i can say, "but mac-gamer.com has a write up on it." :D
     
  20. SpocKirk

    SpocKirk Forum Moderator and Donator

    Joined:
    Feb 28, 2000
    Messages:
    4,813
    Likes Received:
    25
    Market Rating:
    0
Thread Status:
Not open for further replies.

Hitometer: 54,601,131 since 1995